Effectiveness of a Delirium Prevention Bundle in Hospitalized - - PowerPoint PPT Presentation

Evaluating the Feasibility and Effectiveness of a Delirium Prevention Bundle in Hospitalized Critically Ill Patients

Claudia DiSabatino Smith, PhD, RN, NE-BC Petra Grami, BSN, RN, CVRN II

Learning Objectives

• Describe the components of a prevention

bundle that was designed to decrease the incidence and duration of delirium in hospitalized critically ill patients.

• Discuss findings of a study that tested the

use of a delirium prevention bundle in hospitalized critical care patients.

Purpose

The purpose of the research study was to evaluate the effectiveness of a delirium prevention bundle in reducing the incidence of delirium in hospitalized critically ill adult patients.

Research Questions

1. What is the feasibility of Staff RNs effectively assessing patients for delirium using the CAM- ICU? 2. What is the proportion of patients with delirium? 3. Does an intervention bundle prevent or reduce delirium? 4. Are there components of the bundle that are problematic to achieve? 5. Which components of the intervention bundle contribute significantly to reduce or prevent delirium?

Methods

• Design

– Prospective, multiphase – Interventional cohort study

• Setting

– Houston, Texas, USA – Large tertiary care center – Two similar medical-surgical ICUs

• Sample (n=782 intervention study)

– 23,430 observations (8,070 for the intervention unit) - patients have <30 days

• f observations.

Similar Units

Average Case Mix Index

Month Control Unit Intervention Unit

January 2.836 2.707 February 2.894 2.496 March 2.673 3.095 April 3.690 3.783 May 3.271 2.718 June 2.700 3.932 July 2.661 3.657 Total 20.724 22.339

• Avg. CMI

2.960 3.198

Inclusion/Exclusion Criteria

• Critically ill patients

housed in the two study units

• English speaking adult
• Adequate Corrected

Hearing; able to follow commands

• Richmond Agitation-

sedation score (RASS) above -4

• Non-English speaking

patients

• RASS score of -4 or -5
• Patients who have been

in the ICU >4 months

• Lateral transfers:

Patient s transferred from intervention unit to control unit

• Patients in first 12 hours

post surgery who had general anesthesia

• Primary admission

diagnosis new neurological event Exclusion Inclusion

Data Collection Methodology

• Research Coordinator Assistant

– Staff Nurse Documentation

• Evidence-based Researcher-generated

Tools

– Delirium Data Collection Form – Significant Events Form

• Valid & Reliable Instruments

– Richmond Agitation-Sedation Scale (RASS)1 – Confusion Assessment Method for the Intensive Care Unit (CAM-ICU)2

1 Sessler, C.N., Gosnell, M.S., Grap, M.J., et al. (2002). 2 Ely, E.W., Inouye, S.K., Bernard, G.R., Gordon, S., Francis, J., May, L., Truman,

B., Speroff, T., Gautam, S.,Margolin, R., Hart, R.P., Dittus, R. (2001).

The Delirium Prevention Bundle

Prevent or Reduce duration

• f delirium

Sedation Cessation (vents

• nly)

Adequate Pain control Meaningful Sensory stimulation Daily Progressive Mobility Sleep Promotion

Results

7S -1&2 Control Unit 7S-3 Intervention Unit P value

N (%) 512 (65.6) 269 (34.4) Female, n (%) 247 (48.5) 137 (50.9)

0.524

Race, n (%)

0.730

White, non-Hispanic 243 (49.2) 121 (46.7) African-American 175 (35.4) 90 (34.8) Asian 10 (2.0) 4 (1.5) Hispanic 56 (11.3) 38 (14.7) Other 10 (2.0) 6 (2.3) Age Categories, n (%)

0.524

<45 83 (16.2) 44 (16.5) 45-64 206 (40.3) 105 (39.3) 65-74 107 (20.9) 48 (18.0) 75-84 78 (15.3) 53 (19.8) ≥85 37 (7.2) 17 (6.4) # Co morbidities n(%)

0.017

<3 211 (41.2) 89 (33.1) 3-5 211 (41.2) 112 (41.6) >5 90 (17.6) 68 (25.3)

“0” delirium incidents > 1 Delirium Incident p

• No. Days on Mech. Vent,

n (mean)95%CI 527 (0.26) 0.16-0.35 254 (3.76) 3.04- 4.49

<0.0001

• No. Days in Restraint,

n (mean)95%CI 527 (0.06) 0.02-0.11 254 (2.65) 2.08- 3.22

<0.0001

LOS in ICU >3 days, n (%) 134 (25.4) 162 (63.8)

<0.001

• No. Co Morbidities, n (%)

<0.001

<3 232 (44.0) 68 (26.8) 3-5 214 (40.6) 109 (42.9) >5 81 (15.4) 77 (30.3) Female Gender n (%) 263 (50.0) 121 (48.0)

0.604

Age categories n (%)

<0.001

<45 101 (19.3) 26 (10.2) 45-64 225 (42.9) 8 6(33.9) 65-74 94 (17.9) 61 (24.0) 75-84 81 (15.5) 50 (19.7) ≥85 23 (4.39) 31 (12.2) Days on Mech Vent >0, n (%) 63 (12.0) 128 (50.4)

<0.001

Days in Restraint >0, n(%) 16 (3.04) 115 (45.3)

<0.001

Percent change in odds of delirium by day # Subjects (Observa- tions) 7S-1&2 Control Unit % (p-value) 7S-3 Interven- tion Unit % (p-value)

Treatment Effect, % p-value for treatment effect

Unadjusted delirium model (p=0.11) 668 (2687) +2.5 (0.13)

• 4.8

(0.08)

• 7.1

0.021

Delirium model adjusted w age, co- morbidities, LOS variable (p<0.0001) 668 (2687) +1.7 (0.30)

• 6.1

(0.025)

• 7.7

0.012

Gender Male 344 (1319) +6.5 (0.02) +1.4 (0.71)

• 4.8

0.29 Female 321 (1344)

• 0.5

(0.81)

• 11.2

(0.005)

• 10.7

0.016

Age 65-74 139 (621) +10.4 (0.02)

• 13.7

(0.003)

• 21.9

<0.001

Liver Disease No 597 (2326) +4.1 (0.02)

• 4.9

(0.10)

• 8.6

0.009

Yes 65 (328)

• 10.0

(0.06)

• 6.5

(0.36)

+3.9

0.67 Any Kidney Disease No 576 (2328) +2.7 (0.12)

• 3.9

(0.16)

• 6.5

0.041

Yes 86 (326) +0.3 (0.97)

• 22.5

(0.07)

• 22.7

0.086 Co Morbidities <3 246 (928) +1.7 (0.47)

• 11.0

(0.04)

• 12.5

0.031

3-5 277 (1126) +3.8 (0.13) +11.8 (0.02)

+7.7

0.18 >5 145 (633) +1.6 (0.84)

• 14.4

(0.002)

• 15.6

0.052 History of Alcohol No 230 (868) +1.4 (0.61)

• 0.8

(0.90)

• 2.1

0.75 Yes 50 (162) +1.4 (0.92) +64.6 (0.11)

+62.4

0.15 Optimum Conditions: Age 65- 74, female, no liver disease 49 (221) +9.1 (0.002)

• 27.2

(<0.001)

• 33.3

<0.001

Table 14. Results from Longitudinal Logistic Regression Analyses – of Mobility Elements in Intervention Unit Only

Percent change in odds of delirium by day Variables Coding Variable effect @ initial status: (% change in del

• dds, p-value)
• No. Subjects

(Observations) Variable Codes

• No. Subjects (Observations)

Variable effect p-value for variable effect 1

Mobility elements, present? Code: (0,1) = (no, yes) (-89, p=0.044) 236(1016)

• 25.2(0.19)

14(19)

• 4.8(0.093)

222(997) +27.4 0.27 Turning every two hours Code: (0,1) = (no, yes) (-93.4, p=0.007) 235(1006)

• 28.7(0.046)

19(29)

• 4.4(0.13)

216(977) +34.2 0.080 PROM exercises & active ROM per guidelines Code: (0,1) = (no, yes) (-86.7, p=0.049) 223(881)

• 27.4(0.15)

21(27)

• 2.3(0.45)

202(854) +34.5 0.18

Table 15. Results from Longitudinal Logistic Regression Analyses – of Out-of-Bed Elements in Intervention Unit Only

Percent change in odds of delirium by day

Variables Coding Variable effect @ initial status: (% change in del

• dds, p-value)
• No. Subjects

(Observation s) Variable Codes

• No. Subjects

(Observations) Variable effect p-value for variable effect 1

Out-of-Bed elements, present? Code: (0,1) = (no, yes) (-64, p=0.002 ) 235(1002)

• 6.6(0.048)

55(514)

• 0.8(0.84)

180(488) +6.2 0.14 Dangle at edge of bed w/feet planted Code: (0,1) = (no, yes) (+0.1, p>0.99) 234(998)

• 5.0(0.08)

170(899)

• 3.1(0.77)

65(99) +2.1 0.84 Active transfer to chair (20 min TID) Code: (0,1) = (no, yes) (-51.4, p=0.030) 234(1008)

• 6.9(0.034)

40(557)

• 1.5(0.69)

194(451) +5.8 0.17 Ambulation & return to bed (1 to 2 steps per day) Code: (0,1) = (no, yes) (-62.1, p=0.066) 235(1004)

• 5.4(0.057)

174(910) +2.8(0.78) 61(94) +8.7 0.41

Table 10. Results from Longitudinal Logistic Regression Analyses –Adjusted for Effects of Selected Time-dependent Variables using only patients with GTE 2 days of non-missing delirium Percent change in odds of delirium by day Variables No. Subjects (Observati

• ns)

7S-1&2 Control Unit %(p-value) 7S-3 Intervention Unit %(p-value) Treatment Effect, % p-value for treatment effect

Unadjusted model 494(2513) +2.2(0.18)

• 5.4(0.049)
• 7.4

0.017 LTD’s, adj. model OR =4.61, p <0.001 Coded (0,1) = (<3, ≥3) 494(2513) +1.2(0.46)

• 1.6(0.57)
• 2.3

0.38 LTDs <3 442(1578)

• 1.0(0.81)

+2.3(0.60) +3.3 0.58 LTDs ≥3 187(935) +2.3(0.26)

• 8.8(0.040)
• 10.9

0.018 Sedation Protocol, adj. model OR = 1.81, p=0.34 104(493) +6.2(0.023) +2.9(0.57)

• 3.4

0.59 Sed Prot = no 98(464) +6.1(0.024) +2.9(0.574)

• 3.0

0.59 Sed Prot = yes 22(29) +792(0.33)

• 37.5(0.42)
• 92.1

0.31 Sed Prot = NA 473(1961) +2.4(0.39)

• 1.6(0.69)
• 3.9

0.41 Total Complications, adj. model OR = 3.86, p<0.001 Coded (0,1) = (0, >0) 494(2505) +4.0(0.023)

• 4.2(0.13)
• 7.9

0.012

• No. Comps = 0 453(1509) +6.7(0.004)

+3.7(0.53)

• 2.8

0.64

• No. Comps >0

317(996) +3.0(0.37)

• 3.1(0.21)
• 6.8

0.12

Table 10 continued

Hypoactive Symptoms, adj. model OR = 4.53, p <0.001 494(2505) +2.7(0.12)

• 4.3(0.12)
• 6.8

0.03 Any Hypoactive symptoms = no 489(2043) +2.0(0.30)

• 7.5(0.055)
• 9.2

0.027 Any hypoactive symptoms = yes 171(462) +8.7(0.070 ) +1.9(0.62)

• 6.3

0.28 Hyperactive Symptoms, adj. model OR = 7.91, p <0.001 494(2505) +4.3(0.019 )

• 3.0(0.28)
• 7.0

0.028 Any Hyperactive symptoms = no 474(1888) +4.2(0.051 ) +1.2(0.67)

• 2.5

0.55 Any hyperactive symptoms = yes 200(617) +9.8(0.016 )

• 2.4(0.53)
• 11.0

0.03 Mixed Hypo/Hyper, adj. model OR = 2.98, p <0.001 250(854) +10.1(0.00 3) +0.3(0.92)

• 8.9

0.039 Any Hypo/Hyper symptoms = no 233(629) +12.4(0.00 2) +6.2(0.13)

• 5.6

0.29 Any Hypo/Hyper symptoms = yes 96(225) +6.0(0.39)

• 4.7(0.37)
• 10.1

0.23 Confusion, adj. model OR = 21.2, p <0.001 494(2505) +2.8(0.12)

• 4.5(0.11)
• 7.1

0.026 Confusion = no 480(2128) +1.6(0.40)

• 7.5(0.018)
• 9.0

0.012 Confusion = yes 141(377) +6.1(0.20) +14.4(0.039) +7.8 0.35

Table 11. Results from Longitudinal Logistic Regression Analyses –of Selected Time- dependent Variables in Intervention Unit Only, using only patients with GTE 2 days of non- missing delirium Percent change in odds of delirium by day Variables

• No. Subjects

(Observations) Variable Codes Variable effect p-value for variable effect 1 Sensory elements, total no. Code: (0, 1) = (<4, ≥4) OR = 1.77, p = 0.19 175(965)

• 6.2(0.22)
• 5.9(0.048)

+0.3 0.96 Sleep elements, present? Code: (0,1) = (no, yes) OR = 0.29, p = 0.28 174(912)

• 15.3(0.51)
• 4.4(0.13)

+12.8 0.63 Pain controlled Code: (0,1) = (no, yes) OR = 1.78, p = 0.18 173(853)

• 2.8(0.59)
• 3.5(0.26)
• 0.8

0.88 Mobility elements, present? Code: (0, 1) = (no, yes) OR = 0.15, p = 0.086 175(955)

• 22.8(0.25)
• 5.4(0.060)

+22.6 0.36 Out of Bed elements, present? Code: (0,1) = (no, yes) OR = 0.37, p = 0.004 175(942)

• 7.1(0.035)
• 1.6(0.68)

+6.0 0.16

Conclusions

• Patients who are on the mechanical ventilator, who

are restrained, or who have an ICU length of stay >3 days are significantly more likely to develop delirium.

• Delirium is more likely to occur in patients with more

invasive tubes, and in those in which significant hypoxic events occur.

• The intervention bundle significantly reduced the

incidence of delirium in certain critically care patients.

• Some elements of the delirium prevention bundle

were more feasible to implement than other elements.

• Mobility elements show a significant day1 treatment

effect.

Clinical Implications

• It is feasible for staff nurses to effectively use

the CAM-ICU to assess patients for delirium.

• It is feasible to use a delirium prevention

bundle in the ICU setting.

• Hospitals should consider implementing a

core model of delirium prevention care that combines evidence-based strategies with nursing interventions that are integrated into routine ICU care.

Research Implications

• A larger sample size in the intervention unit is

needed to test the effectiveness of some elements (sensory) of the delirium prevention bundle.

• Further research is needed to test a feasible

delirium prevention bundle in a multi-site study.

• Research is needed to test the effectiveness
• f a multidisciplinary consultation team in

treatment of delirium positive critically ill patients.

Limitations

• Hearing aids and glasses not

available for patients

• Lack of physical therapy assistance

early in the study

• Physician-driven weaning method
• Effects of medications not

considered

• Variability in documentation practices

between units

Acknowledgements

• Staff Nurses on 7South3

– Actively involved in the development of the study proposal and researcher-generated data collection tools, and supported and encouraged nursing staff buy-in

• Charles Baimbridge – Statistician

– Analysis & reporting of study data

• Jan Foster, PhD, RN

– Actively involved in the development of the study proposal and staff training in the use of the CAM-ICU

References

• 1 Sessler, C.N., Gosnell, M.S., Grap, M.J., et al.

(2002). The Richmond Agitation-Sedation Scale: Validity and reliability in adult intensive care unit

• patients. American Journal of Respiratory Critical

Care Medicine, 166, 1338-1344.

• 2 Ely, E.W., Inouye, S.K., Bernard, G.R., Gordon,

S., Francis, J., May, L., Truman, B., Speroff, T., Gautam, S.,Margolin, R., Hart, R.P., Dittus, R. (2001). Delirium in mechanically ventilated patients: Validity and reliability of the confusion assessment method for the intensive care unit (CAM-ICU). Journal of the American Medical Association, 286(21), 2703-2710.

Claudia DiSabatino Smith, PhD, RN, NE-BC

• St. Luke’s Hospital, Houston, Texas, USA

csmith1@sleh.com